COST PREDICTION DATASET

Accurate prediction of customer acquisition costs has become a critical business metric, enabling organizations to strategically allocate resources for growth and client acquisition. Customer acquisition cost represents the total expenditure incurred to onboard a new client. However, building an effective prediction model is a complex task due to the presence of numerous independent variables and categorical data with high cardinality, as observed in the Food Mart X dataset and real-world scenarios. Estimating this cost remains a significant challenge for researchers and practitioners alike.
This study proposes a machine learning-based approach to predict customer acquisition costs with high accuracy. Leveraging the Food Mart X dataset, which contains records of 48,000 customers, the proposed models employ algorithms such as Decision Tree (DT), Random Forest (RF), and Bagging Forest (BF) to analyze and predict acquisition costs. Among these methods, the Random Forest algorithm demonstrated the highest accuracy compared to Decision Tree and Bagging methods, showcasing its robustness in handling complex datasets.
While existing literature highlights the potential of neural network models in cost prediction, this study focuses on traditional machine learning techniques. The findings confirm that machine learning methods, particularly Random Forest, can serve as reliable tools for accurate customer acquisition cost prediction, offering valuable insights for businesses to optimize their acquisition strategies.